The Government of Kazakhstan, NVIDIA, and Firebird to Invest $10 Billion in AI: How New Capacities Will Change Fintech

The first phase of the project is estimated at $5 billion, of which $1 billion will be provided by Kazakhtelecom. Commercial launch is scheduled for 2027. In the future, the capacity of the data centers will grow from 300 MW to 1 GW, and the equipment fleet will amount to up to 100,000 NVIDIA GPUs (GB300 and Vera Rubin models).

Country and Market

Kazakhstan is consistently securing its status as a regional digitalization hub in Central Asia. Attracting international giants shows the state’s readiness to build basic infrastructure for complex computing within the country.

For the banking sector, this solves the problem of physical access to capacities. Previously, Finteqstan wrote that major players like Halyk Bank are already implementing AI for employee evaluation; now the market will receive a local base for heavier tasks — from streaming anti-fraud and credit scoring to generative assistants.

Why It Matters and What’s Next

Financial technologies directly depend on data processing speed. Training proprietary language models or complex analytical systems requires enormous server resources, which are expensive for banks to purchase and maintain on their own.

The emergence of a local “Data Center Valley” lowers the barrier to entry into heavy AI development for financial organizations and accelerates the launch of new digital products to the market.

Next, the market will observe the implementation of the first phase by 2027. The successful launch of a 125 MW data center will show how quickly the announced capacities will become available to commercial clients and on what terms banks will be able to rent NVIDIA clusters for their tasks.